Silicon-based agent mitigates fatty liver formation in a CDAHFD60-induced MASH mouse model by enhancing hepatic function.

1
Introduction
Metabolic dysfunction-associated steatohepatitis (MASH) is a severe form of metabolic dysfunction-associated steatotic liver disease (MASLD) characterized by hepatic inflammation and hepatocellular injury in individuals who do not consume excessive amounts of alcohol. Currently, pharmacological treatments for MASH remain insufficient, and lifestyle modifications such as dietary improvement and physical exercise constitute the primary therapeutic approaches. MASH is a progressive disease that can lead to fibrosis, cirrhosis, hepatocellular carcinoma, and ultimately liver failure [1,2]. Owing to the global prevalence of obesity and metabolic syndrome, MASH has emerged as a major public health concern worldwide [3]. Although lifestyle modification remains the cornerstone of therapy, effective pharmacological interventions are urgently needed [4]. Consequently, the development of therapeutic agents that can prevent disease progression and reduce the burden of MASH -related complications has become an urgent priority.
Oxidative stress plays a central role in the pathogenesis of MASH. Excessive lipid accumulation in hepatocytes impairs mitochondrial antioxidative capacity and promotes the production of reactive oxygen species (ROS). The resulting oxidative stress induces cellular damage and stimulates the production of inflammatory cytokines, thereby contributing to MASH development [[5], [6], [7]]. Therefore, suitable antioxidant agents may serve as effective therapeutic options for MASH.
Silicon (Si)–based agent is a novel antioxidant that react with water under neutral to weakly alkaline conditions to continuously generate large amounts of molecular hydrogen, a potent antioxidant [8]. Oral administration of the Si-based agent increases hydrogen generation in the intestinal tract, primarily in the colon, and the agent itself is converted into silicon dioxide after the reaction and subsequently excreted from the body [9]. The Si-based agent represents an effective method for in vivo hydrogen delivery and has been shown to alleviate symptoms in several oxidative stress–related disease models, including Parkinson's disease, ulcerative colitis, and intestinal ischemia–reperfusion injury [[9], [10], [11]].
In this study, we investigated the effects of a Si-based agent on oxidative stress–related pathology in MASH using a choline-deficient, l-amino acid–defined, high-fat diet containing 60 kcal% fat and 0.1% methionine (CDAHFD60)–induced MASH mouse model.

2
Materials & methods

Ethical statement
All experimental procedures involving animals were approved by the Animal Experiment Committee of Osaka University (approval number: 24-085-003) and conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The study was also performed in compliance with the ARRIVE guidelines. Every precaution was taken to minimize both the number of animals used and their suffering. When animals experienced difficulty feeding or drinking, food was placed directly on the bedding, and agar-based jelly was provided as a nutritional supplement. Furthermore, if any signs of distress or humane endpoints were observed—such as impaired feeding or drinking behavior, respiratory difficulties, self-injury, or a rapid body weight loss of 20% or more within a few days—the affected mice were immediately euthanized by intraperitoneal administration of pentobarbital sodium (200 mg/kg).

2.1
Animals
Five-week-old male C57BL/6J mice (n = 26; The Jackson Laboratory Japan, Inc., Kanagawa, Japan) were used in all experiments. The mice were housed under controlled temperature conditions (23–25 °C) with free access to water and a standard rodent diet. The MASH model mice were generated by feeding a CDAHFD60 diet (Oriental Yeast Co., Ltd., Tokyo, Japan) according to previously reported methods [12]. Silicon powder (<45 μm) was sieved and processed into a Si-based agent using a previously described bead-milling method [8]. High-purity Si powder was mechanically milled in ethanol to reduce particle size and improve uniformity, followed by drying. The resulting fine Si particles were then surface-modified by treatment with an aqueous hydrogen peroxide solution to enhance hydrophilicity. After filtration, ethanol washing, and centrifugation, the solid fraction was dried to obtain the final Si-based agent. The Si-based agent administered by mixing it into the CDAHFD60 diet at a concentration of 2.5%, replacing an equal amount of carbohydrate maltodextrin. The experimental schedule is shown in Fig. 1. During the habituation period, all mice were fed AIN93 M (Oriental Yeast Co., Ltd.). After two weeks, the mice were randomly divided into two groups (13 mice per group): an untreated control group (Con group) fed the CDAHFD60 diet and a Si-based agent–treated group (Si group) fed the CDAHFD60 diet containing 2.5% Si-based agent. Throughout the 3-month feeding period, body weight was measured weekly, and blood samples were collected at baseline (0 weeks), at 1.5 months (6 weeks), and at 3 months (12 weeks). At the end of the feeding period, the mice were perfusion-fixed, and both macroscopic and histological analyses of the liver were performed.

2.2
Blood test
Serum biochemical analyses were outsourced to Nagahama Bio Science Laboratory (Oriental Yeast Co., Ltd. Group, Shiga, Japan) and performed using a Hitachi 7180 Clinical Analyzer (Hitachi High-Tech Co., Tokyo, Japan) according to the following methods.•AST (aspartate aminotransferase, IU/L): JSCC-standardized method (ultraviolet–visible absorption spectroscopy; UV–VIS)

•ALT (alanine aminotransferase, IU/L): JSCC-standardized method (ultraviolet–visible absorption spectroscopy; UV–VIS)

•T-BIL (total bilirubin, mg/dL): Enzymatic method using bilirubin oxidase

•TBA (total bile acids, μmol/L): Enzyme cycling method using 3α-hydroxysteroid dehydrogenase

•T-CHO (total cholesterol, mg/dL): Enzymatic method using cholesterol esterase and cholesterol dehydrogenase

•TG (triglycerides, mg/dL): Enzymatic method using lipoprotein lipase and related enzymes

•LDL-C (low-density lipoprotein cholesterol, mg/dL): Direct method based on selective solubilization using surfactants while keeping LDL-C intact

•HDL-C (high-density lipoprotein cholesterol, mg/dL): Direct method based on selective solubilization using surfactants while keeping HDL-C intact

The raw measured values for each sample in the biochemical blood tests are provided in Supplementary Table 1. To normalize intersample variations, comparative analyses were performed using baseline-relative rates of change.

2.3
Tissue section preparation
After perfusion with 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB; pH 7.4) under deep anesthesia using a combination anesthetic (0.3 mg/kg medetomidine, 4.0 mg/kg midazolam, and 5.0 mg/kg butorphanol), the entire liver was removed and fixed in the same PFA solution. Each liver was then divided into two portions: one processed for paraffin-embedded section preparation and the other used for frozen section preparation according to the procedures described below.

2.4
Paraffin-embedded sections
After stepwise dehydration with ethanol (70%, 80%, 90%, and 100%), the liver samples were embedded in paraffin (Tissue Preparation T580, FALMA, Tokyo, Japan) using Clear Plus (FALMA). Paraffin-embedded liver sections, 7 μm thick, were prepared using a microtome (RM2145, Leica Microsystems K.K., Tokyo, Japan) and mounted on MAS-coated glass slides (Matsunami Glass, Osaka, Japan). The slides were stored at 4 °C until use.

2.5
Frozen sections
To prevent ice crystal formation, liver samples were immersed in 0.1 M phosphate buffer (PB) containing 30% sucrose and then rapidly frozen in powdered dry ice. Frozen sections, 20 μm thick, were prepared using a cryostat microtome and mounted on MAS-coated glass slides (Matsunami Glass, Osaka, Japan). The sections were stored at −80 °C until use.

2.6
Morphological analysis
2.6.1
HE staining
After deparaffinization, the sections were stained with hematoxylin solution (FUJIFILM Wako Chemicals Corporation, Osaka, Japan) at 22 ± 2 °C for 5 min, rinsed under running tap water for 10 min, and subsequently stained with eosin solution (FUJIFILM Wako Chemicals Corporation, Osaka, Japan) at 22 ± 2 °C for 3 min. Following staining, all slides were dehydrated through a graded ethanol series and mounted using Entellan (Merck KGaA, Darmstadt, Germany).

2.6.2
Lipid staining (Sudan stain, Oil red O stain)
Frozen sections were thoroughly air-dried, rinsed with running tap water, and immersed in alcohol (Sudan: 1 h; Oil Red O: 30 min) at 37 °C for equilibration. Lipid staining (Muto Pure Chemicals, Tokyo, Japan; Sudan: 50% ethanol; Oil Red O: 60% isopropanol) was then performed, followed by rapid differentiation in alcohol (Sudan: 50% ethanol; Oil Red O: 60% isopropanol). The sections were washed again to adjust color tone and finally mounted using PermaFluor (Thermo Fisher Scientific, Waltham, MA, USA).

2.6.3
Image analysis
All stained samples were imaged, quantified, and analyzed using a Keyence BZ-X700 microscope with the BZ-H3M Measurement Module (Keyence Corporation, Osaka, Japan).
Quantitative values for each staining were calculated as follows.•Hematoxylin and eosin (HE) staining (%): [Unstained vacuolar area/Total liver area] × 100

•Lipid staining (%): [Brown-stained lipid area/Total liver area] × 100

The peritissue spaces, vascular regions, and bile duct areas within the images were excluded from the total liver area. All microscopic images used for image analysis, obtained from 13 mice per group (26 mice in total), are included in Supplementary Figs. 1–3.

2.7
Oxidative stress analysis
Oxidative stress was assessed as previously described [10]. Under deep anesthesia, whole blood was collected from the right atrium of mice in the Con and Si groups, centrifuged (3000 rpm, 10 min, 4 °C), and serum was stored at −80 °C until analysis. Serum reactive oxygen metabolites (dROMs) and antioxidant capacity (BAP) were measured using the REDOXLIBLA system (Wismerll Co., Ltd., Tokyo, Japan). dROMs values were expressed as U.Carr (1 U.Carr = 0.8 mg/L H2O2), and BAP was expressed in μM based on ferric-to-ferrous ion reduction. The BAP/dROMs ratio was used for comparative analysis.

2.8
Statistical analysis
Serum biochemical parameters and body weight were analyzed using a linear mixed-effects model. Group (treatment with Si-based agent vs. control) and Timepoint (0, 6, and 12 weeks) were included as fixed effects, and individual animals were included as a random effect to account for repeated measurements. The model incorporated the main effects of Group and Timepoint as well as their interaction (Group × Timepoint). When a significant interaction was detected, post hoc comparisons between groups at each timepoint were performed using unpaired t-tests.
Quantitative image analysis data for each histological staining were first assessed for normality using the Shapiro–Wilk test and for homogeneity of variances using Levene's test. When both assumptions were satisfied, statistical comparisons were performed using an unpaired t-test. When normality or homogeneity of variance was not met, the Mann–Whitney U test was applied.
Statistical analyses were conducted using JMP Student Edition 18 (SAS Institute Inc., Cary, NC, USA). Statistical significance was defined as follows: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. Data are presented as mean ± standard error of the mean (SEM). All measurement parameters, including mean values and confidence intervals, are summarized in Supplementary Table 2.

3
Results
3.1
The Si-based agent alleviated hepatic dysfunction associated with MASH
Mice were fed either a CDAHFD60 diet containing 2.5% Si-based agent (Si group) or CDAHFD60 diet alone (Con group). Blood samples were collected at three time points—immediately after the start of feeding (0 weeks; 0W), 6 weeks (6W), and 12 weeks (12W)—and subjected to biochemical analyses to assess liver function. Aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels are known to increase during liver injury. To evaluate liver function, serum AST and ALT levels and the AST/ALT ratio were assessed. In both groups, AST and ALT levels markedly increased at 6 weeks after MASH induction compared with baseline values (Fig. 2a and b). Although ALT levels were significantly higher in the Si-treated group at 6 and 12 weeks (51.7% higher than the Con group), no significant differences in the AST/ALT ratio were observed between groups at any time point. Serum transaminases were further analyzed using a linear mixed-effects model with Group and Timepoint as fixed effects and individual animals as a random effect. For AST, only the main effect of Timepoint was significant. In contrast, ALT showed significant main effects of Timepoint and Group, as well as a significant Timepoint × Group interaction (p < 0.0001). Post hoc analyses revealed significant group differences at both 6 and 12 weeks (unpaired t-tests, p < 0.0001). For the AST/ALT ratio, only the main effect of Timepoint was significant, with no significant Group effect or interaction. (Fig. 2c).
Although ALT levels were significantly increased by treatment, AST levels and the AST/ALT ratio remained unchanged, suggesting no apparent exacerbation of hepatocellular injury in Si group.
To further evaluate hepatic metabolic and secretory functions, as well as bile circulation, total bilirubin (T-BIL) and total bile acid (TBA) levels were measured. In the control group, serum total bilirubin (T-BIL) levels remained elevated from 6 weeks after MASH induction, accompanied by a marked increase in total bile acids (TBA) at 6 weeks that further increased at 12 weeks (Fig. 2d and e). In contrast, the Si-treated group showed significant attenuation of T-BIL elevation at 12 weeks (23.6% lower than the Con group), together with significant suppression of TBA accumulation at the same time point (42.2% lower than the Con group). Linear mixed-effects model analysis revealed significant main effects of Timepoint and Group, as well as significant Timepoint × Group interactions for both T-BIL and TBA (p < 0.05). Post hoc analyses demonstrated significant group differences in T-BIL only at 12 weeks (p < 0.001), whereas significant differences in TBA were observed at both 6 weeks (p < 0.05) and 12 weeks (p < 0.0001).
These findings suggest that the Si-based agent primarily improved cholestatic features and bile acid dysregulation rather than hepatocellular injury associated with MASH progression.

3.2
Si-based agent attenuated hepatic steatosis in the MASH mouse model
Next, the liver was subjected to morphological analysis. Gross observation revealed pale pink, glossy livers with fatty changes in both groups, without any remarkable differences between them (Fig. 3a and b). HE staining of paraffin-embedded sections showed numerous vacuoles in the hepatic parenchyma of the Con group. In addition, black granular deposits suggestive of iron deposition and bile pigment accumulation were detected (Fig. 3c). In contrast, although vacuoles and black deposits were also present in the Si group, their abundance was lower than that in the Con group (Fig. 3d). Quantitative analysis of the vacuolated area relative to the total hepatic parenchyma revealed that the proportion of vacuolated regions was significantly smaller in the Si group than in the Con group (p < 0.0001; 47.1% lower than the Con group; Fig. 3e).
Because the vacuoles observed in HE-stained sections were likely artifacts caused by fat dissolution during paraffin embedding, two types of lipid staining were performed. Sudan staining revealed lipid droplets of various sizes in both groups, but smaller droplets were more frequently observed in the Si group than in the Con group (Fig. 4a and b). Statistical analysis showed that the area occupied by lipid droplets was significantly smaller in the Si group than in the Con group (p < 0.01; 15.5% lower than the Con group; Fig. 4e). Furthermore, analysis of neutral lipids using Oil Red O staining showed that both groups contained neutral lipids; however, smaller lipid droplets were more frequently detected in the Si group (Fig. 4c and d). Consistently, quantitative analysis revealed that the area occupied by neutral lipids was also significantly smaller in the Si group than in the Con group (p < 0.0001; 25.3% lower than the Con group; Fig. 4f).
Taken together, these findings suggest that the Si-based agent effectively attenuated hepatic steatosis in the MASH mouse model.

3.3
The Si-based agent mitigated the reductions in blood lipid levels and body weight induced by CDAHFD60
CDAHFD60-induced MASH mouse models differ from human MASH associated with metabolic syndrome, as they rarely develop obesity or insulin resistance. Instead, choline deficiency causes lipid metabolic abnormalities and body weight loss. To evaluate the efficacy of the Si-based agent in addressing these distinctive pathophysiological features, blood biochemical analyses were performed. In the control group, total cholesterol (T-CHO), triglycerides (TG), LDL-C, and HDL-C markedly decreased at 6 weeks after MASH induction and remained low at 12 weeks (Fig. 5a–d). These decreases were attenuated in the Si-treated group. In particular, significant suppression of T-CHO and HDL-C reductions was observed at both 6 and 12 weeks (T-CHO: 22.2% and HDL-C: 72.7% higher than the Con group). Similarly, TG levels in the Si group also showed attenuation of the decreases at both time points (46.2% higher than the Con group). Notably, LDL-C showed partial recovery at 12 weeks (67.3% higher than the Con group) (Fig. 5c). Linear mixed-effects model analysis demonstrated significant main effects of Timepoint and Group, as well as significant Timepoint × Group interactions, for T-CHO and HDL-C (p < 0.05). Post hoc analyses revealed significant group differences at both 6 weeks (T-CHO, p < 0.001; HDL-C, p < 0.0001) and 12 weeks (T-CHO, p < 0.01; HDL-C, p < 0.05). In contrast, TG showed significant main effects of Timepoint and Group only, while LDL-C exhibited a significant main effect of Timepoint alone (p < 0.01).
Furthermore, changes in body weight were examined. In the Con group, body weight decreased immediately after CDAHFD60 initiation and declined until week 3, followed by a gradual increase with normal growth (Fig. 5e). In contrast, the Si group showed an initial decrease without further decline, and body weight remained higher from week 4 onward compared with the Con group (4-7% higher than the Con group). Mixed-effects model analysis revealed significant main effects of Timepoint and Group (p < 0.0001).
These results suggest that the Si-based agent significantly mitigated body weight loss by alleviating lipid metabolic abnormalities induced by CDAHFD60.

3.4
The Si-based agent alleviated systemic oxidative stress in the CDAHFD60-induced MASH mouse model
Finally, to evaluate whether the Si-based agent exerts antioxidant effects in the CDAHFD60-induced MASH mouse model, serum d-ROMs and BAP assays were performed. Based on previous reports, d-ROMs levels in the Con group were elevated compared with those in normal C57BL/6J mice (d-ROMs:36.0 ± 1.73) [13]; in contrast, this MASH-associated increase was significantly suppressed in the Si group (12% lower than the Con group; Fig. 6a). No significant difference in BAP levels was observed between the two groups (Fig. 6b). The BAP/d-ROMs ratio, which inversely reflects oxidative stress burden, was significantly higher in the Si group than in the Con group (18.7% higher than the Con group; Fig. 6c).
These results indicate that the Si-based agent alleviates systemic oxidative stress by suppressing the MASH-associated increase in circulating hydroperoxides.

4
Discussion
MASH is a form of hepatitis caused by the accumulation of fat in the liver due to nonalcohol-related factors such as overeating and physical inactivity. Persistent MASH increases the risk of liver cirrhosis and hepatocellular carcinoma, emphasizing the importance of early intervention. In the present study, administration of an Si–based agent to MASH mouse models alleviated hepatic fat accumulation (32.9% improvement) and hepatic dysfunction (29.3% improvement)—two major hallmarks of the disease. These findings suggest that the Si-based agent holds promise as a potential therapeutic candidate for MASH.
Oxidative stress plays a crucial role in the pathogenesis of MASLD, the precursor condition of MASH [[5], [6], [7]]. Excessive lipid accumulation in hepatocytes impairs mitochondrial antioxidative capacity, leading to increased production of ROS. Elevated ROS not only induce lipid peroxidation and mitochondrial dysfunction [14] but also activate Kupffer cells and hepatic stellate cells, initiating inflammatory pathways that promote both inflammation and fibrosis [15]. Consequently, hepatocellular injury and fibrosis develop. Thus, oxidative stress is a key driver of the progression from simple steatosis to MASH [16]. Indeed, investigations of oxidative stress markers in patients with MASH have shown that elevated levels of lipid peroxidation products and decreased antioxidant capacity correlate with disease severity [15]. Given the pivotal role of oxidative stress in the onset and progression of MASH, appropriate antioxidant agents may serve as effective therapeutic options. Several antioxidants, including N-acetylcysteine [17], anthocyanins [18], flavonoids [19], and manganese–selenium complexes that mimic superoxide dismutase and catalase activity [20], have been reported to alleviate MASH pathology by reducing oxidative stress.
Furthermore, clinical studies have demonstrated the efficacy of vitamin C and vitamin E supplementation in patients with MASH [21,22]. In addition, antioxidants that selectively scavenge harmful ROS [23], as well as the administration of hydrogen-rich water, have also shown beneficial effects in MASH [[24], [25], [26]], and clinical trials have reported favorable outcomes in patients with MASLD [27]. The Si–based agent generates molecular hydrogen in the intestine and has previously been shown to alleviate symptoms in various oxidative stress–related disease models, including ulcerative colitis, Parkinson's disease, intestinal ischemia–reperfusion injury, and interstitial pneumonia. In the present MASH model, the Si-based agent also attenuated systemic oxidative stress, as evidenced by suppression of circulating lipid peroxides and an increased BAP/d-ROMs ratio (Fig. 6). Given its suitability for in vivo administration, the Si-based agent is expected to mitigate hepatic steatosis. Previous studies utilizing hydrogen-based therapies for MASH have primarily employed high-fat diet (HFD) or choline-deficient, amino acid–defined (CDAA) diets to induce the disease. In contrast, the present study used a CDAHFD60 diet. Histopathological analyses have shown that CDAHFD60 induces more severe hepatic fibrosis than either the HFD or CDAA diet [28]. Therefore, the present findings not only reinforce the therapeutic potential of hydrogen in MASH but also suggest that the Si-based agent represents an effective approach for in vivo hydrogen delivery.
Although the CDAHFD60-induced MASH mouse model is valuable for elucidating the mechanisms of hepatitis and fibrosis, it differs from clinical MASH in that it does not exhibit obesity or insulin resistance [29,30]. Because this model lacks choline, very low-density lipoprotein particles cannot form, preventing triglycerides and cholesterol synthesized in the liver from being secreted into the circulation. Furthermore, because CDAHFD60 is a high-fat diet, lipids progressively accumulate in the liver. Consequently, hepatic lipid accumulation increases while circulating lipid levels decrease, ultimately resulting in body weight loss.
Previous studies have reported that PPAR agonists, which improve hepatic lipid accumulation, inflammation, and fibrosis, do not affect the lean and hypolipidemic phenotype characteristic of the CDAHFD60-induced MASH model [29]. In addition, recent comparative and profiling analyses have demonstrated that when MASH is induced by CDAHFD-based diets (e.g., CDAA or CDAHFD) and subsequently switched to a normal chow diet, body weight and serum cholesterol and triglyceride (TG) levels significantly recover [31]. Although certain pharmacological interventions (e.g., PPAR agonists and farnesoid X receptor agonists) differentially affect fibrosis and inflammation [29,32], recovery of body weight and blood lipid levels generally requires dietary modifications, such as increasing methionine content to ≥0.2%, supplementing with choline, or implementing a chow-reversal protocol.
Interestingly, the Si–based agent not only alleviated hepatic steatosis but also mitigated the hypolipidemia and body weight loss characteristic of the CDAHFD60-induced MASH model. To date, no pharmacological agents effective against hepatic steatosis or steatohepatitis have been reported to improve either hypolipidemia or body weight loss. Therefore, the Si-based agent may not only attenuate liver dysfunction but also enhance systemic lipid metabolism in this model by restoring and regulating hepatic function. However, because the Si-based agent was administered as part of the diet, the possibility that it influenced MASH induction cannot be excluded, representing a limitation of this study.
In the present study, the Si-based agent mitigated hepatic steatosis by alleviating liver dysfunction associated with MASH progression. Unlike previously investigated therapeutic candidates, it also tended to improve systemic lipid metabolism. Given that neither the Si-based agent nor the molecular hydrogen it generates has been associated with adverse effects, this agent holds promise as a novel and safe therapeutic candidate for MASH.

5
Limitations
A limitation of this study is the use of the CDAHFD60-induced MASH mouse model, which does not fully reproduce the metabolic features of human MASH, such as obesity and insulin resistance [12]. Although this model reliably induces hepatic steatosis, inflammation, and fibrosis, these pathological changes occur independently of systemic metabolic dysfunction. Therefore, the therapeutic effects of the Si-based agent observed in this study should be extrapolated to clinical MASH with caution [4]. Further validation using models that incorporate metabolic abnormalities, as well as clinical studies, will be required to clarify the translational relevance of this agent.
Second limitation of this study is that oxidative stress was assessed only systemically using serum-based d-ROMs and BAP assays, and not directly in liver tissue. Although these markers reflect whole-body oxidative stress, they may not accurately represent oxidative stress at the site of liver injury. Therefore, the contribution of hepatic oxidative stress to the protective effects of the Si-based agent remains unclear and warrants further investigation.
Third, food intake and physical activity were not directly assessed in this study and may represent potential confounding factors. Accurate measurement of food intake was difficult because a powdered diet was used. Although previous studies using tablet diets have reported no differences in food intake irrespective of Si-based agent administration [9], this possibility cannot be completely excluded. Therefore, the lack of direct assessment of these parameters should be considered a limitation.

Author contributions
Yo.K designed the study, analyzed the data, and wrote the paper. Yo.K and I.H. performed morphological analysis. Yu.K. and H.K. developed the method for fabrication of Si-based agent. S.S. supervised this study and provided intellectual directions. All authors discussed the findings and commented on this manuscript.

Funding details
This work was supported by Center of Innovation Program (COI Program) Grant Number JPMJCE1310, JST Japan, KAKENHI (No. 23K07436), JSPS, Japan, and the Tohmonkai Foundation, Japan.

Declaration of competing interest
The authors report no conflict of interest.